ud exp eda

Lets review MLB draft season.

Start by loading the packages

suppressMessages({
  library(tidyverse) #ggplot2 dplyr tibble tidyr purrr forecats 
  library(ggrepel) #automatically position non-overlapping text labels
  library(glue) #interpreted literal strings
  library(gt)
  library(gtExtras)
  library(paletteer)
  library(mlbplotR)
})

Next we load the team logos

teams_colors_logos <- mlbplotR::load_mlb_teams() %>% 
  filter(!team_abbr %in% c("AL", "NL", "MLB")) %>% 
  mutate(
    a = rep(1:6, 5), 
    b = sort(rep(1:5, 6), decreasing=T), 
    alpha = ifelse(grepl("A", team_abbr),1,0.75),
    color = ifelse(grepl("E", team_abbr), "b/w", NA)
  )

Now lets load our draft data and clean it up.

exposure <- read.csv("./data/exposure_mar30.csv")

exposure <- exposure %>% 
  mutate(Picked.At = as.Date(as.POSIXct(exposure$Picked.At, format="%Y-%m-%d %H:%M:%S", tz="UTC")), 
         name = paste(First.Name, Last.Name)) %>% 
  select(name, Team, Position, Picked.At, Pick.Number, Draft) %>% 
  left_join(teams_colors_logos %>% select(team_abbr, team_logo_espn), by=c('Team'='team_abbr')) %>% 
  left_join(read.csv("./data/playerids.csv"), by=c('name'='Name')) %>% 
  mutate(
    copy = paste0(name, Draft),
    playerid = as.double(playerid)) %>% 
  distinct(copy, .keep_all = T) %>% 
  left_join(mlbplotR::load_headshots() %>% select(fangraphs_id, espn_headshot) %>% drop_na(fangraphs_id), by=c("playerid"="fangraphs_id")) %>% 
  select(-copy) %>% 
  left_join(read.csv("./projections_season/rankings_mar20.csv") %>% 
              mutate(name = paste(firstName, lastName), 
                     adp = as.numeric(adp)) %>% 
              select(name, adp, projectedPoints, positionRank),
            by=c("name")) %>% 
  mutate(value = Pick.Number-adp, 
         rel_value = round(value/adp, digits = 2),
         positionGroup = gsub("[^A-Z]","",positionRank)) %>% 
  drop_na(adp) %>% 
  arrange(Pick.Number)

## Warning in mask$eval_all_mutate(quo): NAs introduced by coercion

Highest owned players

#group by player
exposure %>% 
  group_by(name, espn_headshot) %>% 
  summarise(count=n()) %>% 
  arrange(-count) %>% 
  ungroup() %>% 
  mutate(own = round(count/length(unique(exposure$Draft)),digits = 2)) %>% 
  slice_head(n=10) %>% 
  gt() %>% 
  gt_img_rows(columns = espn_headshot, height = 50) %>% 
  gt_theme_dark()

## `summarise()` has grouped output by 'name'. You can override using the
## `.groups` argument.

name	espn_headshot	count	own
Mark Canha		20	0.28
J.D. Martinez		19	0.26
Willson Contreras		18	0.25
Jake Cronenworth		17	0.24
Lourdes Gurriel Jr.		17	0.24
Tommy Edman		15	0.21
Aaron Nola		14	0.19
Austin Hays		14	0.19
Justin Turner		14	0.19
Alek Manoah		13	0.18

Lets look at the drafts grouped by date

#drafts by date
drafts_by_date <- exposure %>% 
  group_by(Picked.At) %>% 
  summarize(total_picks = n(),
            total_value = sum(value, na.rm = T), 
            total_rel_value = sum(rel_value, na.rm = T)) %>% 
  mutate(value_per_pick = round(total_value/total_picks,digits = 2), 
         rel_value_per_pick = round(total_rel_value/total_picks,digits=2))

drafts_by_date %>% gt() %>% gt_theme_dark()

Picked.At	total_picks	total_value	total_rel_value	value_per_pick	rel_value_per_pick
2023-02-19	20	-21.0	0.25	-1.05	0.01
2023-02-21	172	-399.5	-0.65	-2.32	0.00
2023-02-22	111	232.2	7.02	2.09	0.06
2023-02-23	177	-26.8	0.50	-0.15	0.00
2023-02-25	120	118.6	3.09	0.99	0.03
2023-02-26	20	-29.3	0.04	-1.46	0.00
2023-02-27	39	-99.8	0.69	-2.56	0.02
2023-02-28	40	1.4	0.11	0.04	0.00
2023-03-01	20	35.1	0.78	1.76	0.04
2023-03-02	20	-13.2	0.10	-0.66	0.00
2023-03-04	40	29.2	0.56	0.73	0.01
2023-03-05	20	107.0	1.36	5.35	0.07
2023-03-06	20	32.6	-0.33	1.63	-0.02
2023-03-08	80	-27.1	0.22	-0.34	0.00
2023-03-09	20	-42.3	0.92	-2.11	0.05
2023-03-11	7	21.0	0.50	3.00	0.07
2023-03-12	57	67.8	0.56	1.19	0.01
2023-03-13	16	-84.2	-0.50	-5.26	-0.03
2023-03-19	80	239.2	2.20	2.99	0.03
2023-03-20	40	-32.9	-0.07	-0.82	0.00
2023-03-26	160	25.3	3.13	0.16	0.02
2023-03-27	20	-167.4	-1.04	-8.37	-0.05
2023-03-29	80	-218.2	-0.47	-2.73	-0.01
2023-03-30	59	-279.9	-0.70	-4.74	-0.01

Top 10 picks from all drafts in terms of value

#top ten picks by value
exposure %>% 
  select(name, team_logo_espn, espn_headshot, Pick.Number, adp, value, rel_value, Picked.At) %>% 
  arrange(-rel_value) %>% 
  slice_head(n=10) %>% 
  
  gt() %>% 
  gt_img_rows(columns = "team_logo_espn", height = 50) %>% 
  gt_img_rows(columns = "espn_headshot", height = 50) %>% 
  gt_theme_dark()

name	team_logo_espn	espn_headshot	Pick.Number	adp	value	rel_value	Picked.At
Aaron Judge			4	1.2	2.8	2.33	2023-02-22
Juan Soto			6	2.5	3.5	1.40	2023-03-26
Ronald Acuña Jr.			6	2.9	3.1	1.07	2023-03-09
Aaron Judge			2	1.2	0.8	0.67	2023-03-30
Aaron Judge			2	1.2	0.8	0.67	2023-02-21
Juan Soto			4	2.5	1.5	0.60	2023-02-27
Julio Rodríguez			8	5.4	2.6	0.48	2023-02-22
Jarred Kelenic			225	153.5	71.5	0.47	2023-02-23
Shohei Ohtani			7	4.8	2.2	0.46	2023-02-22
Jarred Kelenic			221	153.5	67.5	0.44	2023-03-04

team drafted

#group by team drafted
exposure %>% 
  group_by(Team, team_logo_espn) %>% 
  summarise(count=n()) %>% 
  arrange(-count) %>% 
  ungroup() %>% 
  rename(team = team_logo_espn) %>% 
  slice_head(n=10) %>% 
  gt() %>% 
  gt_img_rows(columns = team) %>% 
  gt_theme_dark() 

## `summarise()` has grouped output by 'Team'. You can override using the
## `.groups` argument.

Team	team	count
STL		116
NYM		95
SD		89
LAD		85
ATL		79
LAA		73
TOR		68
NYY		56
SEA		56
HOU		55

by position

#group by position
exposure %>% 
  group_by(Position) %>% 
  summarise(count=n()) %>% 
  arrange(-count) %>% 
  ungroup() %>% 
  mutate(own = round(count/sum(count),digits = 2)) %>% 
  gt() %>% 
  gt_theme_dark()

Position	count	own
SP	430	0.30
RF	171	0.12
LF	165	0.11
3B	162	0.11
1B	114	0.08
CF	102	0.07
SS	102	0.07
2B	94	0.07
C	59	0.04
DH	39	0.03

Number of stacked batters by team

#Number of stacked batters by team
exposure %>% 
  group_by(Draft, Team, team_logo_espn) %>% 
  summarise(count=n()) %>% 
  arrange(-count) %>% 
  ungroup() %>% 
  group_by(Team, team_logo_espn) %>% 
  summarise(count=n()) %>% 
  arrange(-count) %>% 
  gt() %>% 
  gt_img_rows(columns = team_logo_espn) %>% 
  gt_theme_dark()

## `summarise()` has grouped output by 'Draft', 'Team'. You can override using the
## `.groups` argument.
## `summarise()` has grouped output by 'Team'. You can override using the
## `.groups` argument.

team_logo_espn	count
STL
	47
NYM
	44
ATL
	41
LAD
	41
SD
	40
TOR
	38
LAA
	36
MIL
	35
HOU
	34
TB
	34
SEA
	33
SF
	33
MIN
	32
PHI
	32
BOS
	29
CLE
	29
NYY
	29
BAL
	26
ARI
	25
MIA
	24
CWS
	23
KC
	23
TEX
	23
COL
	18
CHC
	17
PIT
	17
DET
	12
CIN
	10
WSH
	2
OAK
	1

Creating objects to merge to the drafts dataframe

ord <- c("P","IF","OF")

#configurations of P, OF, IF
exposure_config <- exposure %>% 
  group_by(Draft, positionGroup) %>% 
  summarise(count=n()) %>% 
  arrange(Draft, factor(positionGroup, levels = ord)) %>% 
  ungroup() %>% 
  group_by(Draft) %>% 
  summarise(config = as.numeric(paste0(count, collapse = ""))) %>% 
  ungroup() 

## `summarise()` has grouped output by 'Draft'. You can override using the
## `.groups` argument.

exposure_config %>% 
  group_by(config) %>% 
  summarise(count = n()) %>% 
  arrange(-count)

## # A tibble: 9 × 2
##   config count
##    <dbl> <int>
## 1    677    27
## 2    686    18
## 3    587    11
## 4    776     8
## 5    578     3
## 6    767     2
## 7    586     1
## 8    676     1
## 9    965     1

#stacked batters by draft
exposure_batters <- exposure %>% 
  filter(positionGroup != "P") %>% 
  group_by(Draft, Team) %>% 
  summarise(batters = n()) %>% 
  ungroup() %>% 
  filter(batters > 1) %>% 
  group_by(Draft) %>% 
  summarise(batters = sum(batters)) %>% 
  arrange(-batters)

## `summarise()` has grouped output by 'Draft'. You can override using the
## `.groups` argument.

# biggest stack per draft
exposure_big_stack <- exposure %>% 
  filter(positionGroup != "P") %>% 
  group_by(Draft, Team) %>% 
  summarise(batters = n()) %>% 
  ungroup() %>% 
  group_by(Draft) %>% 
  summarise(big_stack = max(batters)) %>% 
  ungroup()

## `summarise()` has grouped output by 'Draft'. You can override using the
## `.groups` argument.

#number of teams with stacked batters per draft
exposure_num_teams <- exposure %>% 
  filter(positionGroup != "P") %>% 
  group_by(Draft, Team) %>% 
  summarise(batters = n()) %>% 
  ungroup() %>% 
  filter(batters > 1) %>% 
  group_by(Draft) %>% 
  summarise(teams_stacked = n())

## `summarise()` has grouped output by 'Draft'. You can override using the
## `.groups` argument.

#find first pick of each draft
first_pick <- exposure %>% 
  filter(Pick.Number < 13) %>% select(name, Draft) %>% 
  rename("first_pick" = "name")

#drafts
drafts <- exposure %>% 
  #drop_na() %>% 
  group_by(Draft) %>% 
  summarize(total_picks = n(),
            total_value = sum(value), 
            total_rel_value = sum(rel_value), 
            Picked.At = last(Picked.At)) %>% 
  mutate(value_per_pick = round(total_value/total_picks, digits = 2),
         rel_value_per_pick = round(total_rel_value/total_picks, digits = 2)) %>% 
  arrange(-rel_value_per_pick) %>% 
  left_join(exposure_config, by=c("Draft")) %>% 
  left_join(exposure_batters, by=c("Draft")) %>% 
  left_join(exposure_big_stack, by=c("Draft")) %>% 
  left_join(exposure_num_teams, by=c("Draft")) %>% 
  left_join(first_pick, by=c("Draft")) %>% 
  mutate(file = paste(Picked.At,config, teams_stacked, batters, big_stack, first_pick))

drafts %>% 
  select(file, total_value, total_rel_value, Picked.At) %>% 
  gt()

file	total_value	total_rel_value	Picked.At
2023-02-22 587 5 11 3 Aaron Judge	147.1	3.58000e+00	2023-02-22
2023-02-22 686 3 10 4 Mike Trout	33.8	1.54000e+00	2023-02-22
2023-02-25 686 4 10 3 Shohei Ohtani	120.5	1.62000e+00	2023-02-25
2023-03-05 686 2 8 5 Mookie Betts	107.0	1.36000e+00	2023-03-05
2023-03-26 776 3 11 4 Juan Soto	-7.6	1.37000e+00	2023-03-26
2023-02-22 776 3 7 3 Julio Rodríguez	46.1	1.18000e+00	2023-02-22
2023-02-23 677 4 11 3 Mookie Betts	112.4	1.18000e+00	2023-02-23
2023-03-26 677 3 9 4 Mookie Betts	98.8	1.10000e+00	2023-03-26
2023-03-08 686 6 13 3 Shohei Ohtani	33.3	9.90000e-01	2023-03-08
2023-03-09 587 6 15 3 Ronald Acuña Jr.	-42.3	9.20000e-01	2023-03-09
2023-03-19 677 3 8 4 Juan Soto	140.7	1.03000e+00	2023-03-19
2023-03-01 578 5 14 4 José Ramírez	35.1	7.80000e-01	2023-03-01
2023-03-04 677 5 13 4 Julio Rodríguez	70.6	7.90000e-01	2023-03-04
2023-03-29 677 5 12 4 Mike Trout	27.1	9.00000e-01	2023-03-29
2023-02-22 587 4 10 3 Shohei Ohtani	49.6	9.00000e-01	2023-02-22
2023-02-25 677 4 10 3 Mookie Betts	57.0	8.40000e-01	2023-02-25
2023-03-26 686 4 10 3 Yordan Alvarez	-6.7	7.70000e-01	2023-03-26
2023-02-21 776 4 9 3 Aaron Judge	-50.4	5.90000e-01	2023-02-21
2023-02-23 677 4 11 5 Vladimir Guerrero Jr.	67.6	6.40000e-01	2023-02-23
2023-02-25 677 4 9 3 Mike Trout	-3.2	6.40000e-01	2023-02-25
2023-03-12 686 6 13 3 Aaron Judge	65.3	5.80000e-01	2023-03-12
2023-02-27 586 4 11 4 Juan Soto	-81.9	6.30000e-01	2023-02-27
2023-03-26 677 4 12 4 Julio Rodríguez	44.1	3.50000e-01	2023-03-26
2023-02-23 686 3 12 5 Mookie Betts	63.5	3.60000e-01	2023-02-23
2023-02-23 677 4 12 4 José Ramírez	9.5	4.70000e-01	2023-02-23
2023-03-19 677 4 11 4 Trea Turner	45.3	4.50000e-01	2023-03-19
2023-03-19 677 4 9 3 Shohei Ohtani	69.6	4.80000e-01	2023-03-19
2023-03-29 677 3 11 5 Mike Trout	5.8	3.20000e-01	2023-03-29
2023-03-12 686 3 12 6 Shohei Ohtani	21.1	4.70000e-01	2023-03-12
2023-03-12 686 4 14 4 Shohei Ohtani	10.7	3.00000e-01	2023-03-12
2023-03-19 686 3 10 4 Julio Rodríguez	-16.4	2.40000e-01	2023-03-19
2023-02-21 776 3 6 2 Trea Turner	-87.0	2.10000e-01	2023-02-21
2023-02-19 965 2 4 2 Kyle Tucker	-21.0	2.50000e-01	2023-02-19
2023-02-21 677 2 4 2 Mike Trout	-26.7	2.80000e-01	2023-02-21
2023-02-25 677 4 11 4 Aaron Judge	7.0	2.00000e-01	2023-02-25
2023-02-23 677 5 13 3 José Ramírez	-4.7	2.30000e-01	2023-02-23
2023-02-28 677 2 9 7 José Ramírez	-10.5	2.10000e-01	2023-02-28
2023-02-21 776 1 6 6 Mike Trout	0.9	-7.00000e-02	2023-02-21
2023-03-29 776 4 10 4 Juan Soto	-58.7	5.20417e-18	2023-03-29
2023-02-26 677 4 11 3 José Ramírez	-29.3	4.00000e-02	2023-02-26
2023-03-26 686 3 8 4 Shohei Ohtani	0.2	3.00000e-02	2023-03-26
2023-02-22 767 4 9 3 Vladimir Guerrero Jr.	-61.9	-8.00000e-02	2023-02-22
2023-02-21 776 1 3 3 Kyle Tucker	21.3	8.00000e-02	2023-02-21
2023-03-02 677 5 13 5 Yordan Alvarez	-13.2	1.00000e-01	2023-03-02
2023-03-20 587 3 12 6 Juan Soto	-20.5	9.00000e-02	2023-03-20
2023-02-23 587 3 11 4 Mike Trout	-25.1	-1.00000e-02	2023-02-23
2023-03-08 587 4 11 5 Mookie Betts	-0.6	-9.00000e-02	2023-03-08
2023-03-30 676 4 11 4 Aaron Judge	-92.8	3.00000e-02	2023-03-30
2023-03-26 677 4 11 4 Vladimir Guerrero Jr.	-31.2	8.00000e-02	2023-03-26
2023-02-27 587 4 12 5 Juan Soto	-17.9	6.00000e-02	2023-02-27
2023-02-28 677 3 9 5 José Ramírez	11.9	-1.00000e-01	2023-02-28
2023-02-25 587 4 12 4 Shohei Ohtani	-27.3	6.00000e-02	2023-02-25
2023-02-25 677 4 12 4 Shohei Ohtani	-35.4	-2.70000e-01	2023-02-25
2023-02-21 767 3 6 2 Kyle Tucker	-51.7	-1.90000e-01	2023-02-21
2023-03-04 587 5 14 5 Aaron Judge	-41.4	-2.30000e-01	2023-03-04
2023-03-30 686 5 13 4 Trea Turner	-80.6	-1.10000e-01	2023-03-30
2023-03-20 677 3 12 5 José Ramírez	-12.4	-1.60000e-01	2023-03-20
2023-03-26 587 6 14 3 Mike Trout	-17.6	-1.40000e-01	2023-03-26
2023-02-21 776 3 7 3 Shohei Ohtani	-13.2	-3.60000e-01	2023-02-21
2023-03-06 686 4 11 3 Ronald Acuña Jr.	32.6	-3.30000e-01	2023-03-06
2023-02-22 677 3 10 5 José Ramírez	-76.0	-4.10000e-01	2023-02-22
2023-03-08 686 4 12 4 Ronald Acuña Jr.	-32.5	-3.80000e-01	2023-03-08
2023-03-08 686 4 11 4 Juan Soto	-27.3	-3.00000e-01	2023-03-08
2023-03-26 686 3 8 4 Mookie Betts	-54.7	-4.30000e-01	2023-03-26
2023-03-30 686 3 11 7 Yordan Alvarez	-106.5	-6.20000e-01	2023-03-30
2023-02-23 587 4 11 4 Mike Trout	-82.5	-5.10000e-01	2023-02-23
2023-02-23 578 3 8 4 Juan Soto	-77.8	-7.00000e-01	2023-02-23
2023-03-13 677 5 14 3 Mike Trout	-92.5	-7.90000e-01	2023-03-13
2023-03-27 677 4 8 2 Vladimir Guerrero Jr.	-167.4	-1.04000e+00	2023-03-27
2023-02-22 686 3 6 2 Gerrit Cole	-101.8	-9.10000e-01	2023-02-22
2023-02-23 578 3 12 5 Yordan Alvarez	-87.1	-1.13000e+00	2023-02-23
2023-03-29 677 4 12 4 Paul Goldschmidt	-192.4	-1.69000e+00	2023-03-29

#create a list from the dataframe
exp_list <- split(exposure, exposure$Draft)

name_mapping <- data.frame(
  old_names = drafts$Draft, 
  new_names = drafts$file, 
  stringsAsFactors = F
)

# Find the indices of the old names in the dataframe
name_indices <- match(names(exp_list), name_mapping$old_names)

# Rename the list elements using the new names from the dataframe
names(exp_list)[!is.na(name_indices)] <- name_mapping$new_names[name_indices[!is.na(name_indices)]]